"monitor...using SNMP" is a very broad term. Depending on what MIB Object IDs you query and how, the potential impact can vary widely. You can, for example, query a single variable or hundred (say the ifInOctets for every interface on a large core switch).
If you have an SNMP tool already, I'd rely on direct measurement to get the numbers you are asking for.
If not, then use something like Paessler's free SNMP Tester - download and install it and then point it at a target device and query it for the typical set of OIDs.
https://www.paessler.com/tools/snmptester
In either case measure the data transferred with a tcpdump or Wireshark packet capture.
Also, 60 seconds is a bit aggressive of a cycle for SNMP. We usually see interfaces more like 5-10 minutes or even more. For near real time availability we might use a background ping test that's running more frequently and will flag unreachability quickly.
